Plaintiff-Appellant Brenda Estrella appeals from a judgment of the United States District Court for the Eastern District of New York (Kuntz, J. ), affirming the Commissioner of Social Security's ("Commissioner") denial of disability benefits under Titles II and XVI of the Social Security Act ("SSA" or "the Act"), 42 U.S.C. §§ 401 - 434 and 1381 - 1383. Estrella contends that the administrative law judge ("ALJ") who reviewed her claim, and whose decision the Commissioner adopted, erred by (A) failing to give controlling weight to the opinion of her treating physician and (B) failing to explain the reasons for giving that opinion minimal weight. For the reasons that follow, we vacate the judgment of the district court and remand the case to the Commissioner for further proceedings consistent with this opinion.

*94 BACKGROUND

Estrella worked as an administrative clerk from 1994 until 2008. Beginning in 2002 and continuing through at least 2013, Estrella suffered from, as relevant to this appeal, major depressive disorder, bipolar disorder, and Attention Deficit Hyperactive Disorder. From 2003 to 2006, and again from 2010 to 2013, Estrella took various medications to treat her mental illness and attended monthly psychotherapy sessions. Estrella testified that she did not seek treatment between 2006 and 2010 because she "was in an abusive relationship[,] ... was suffering from severe depression," and had consequently "withdr[awn] [her]self." Admin. R. 108.

In June 2012, Estrella applied for Social Security Disability Insurance and Supplemental Security Income benefits. She claimed that she had been unable to work since August 27, 2008 because of depression, herniated discs, knee pain, diabetes, nerve damage, sleep apnea, and "spasms." Id. at 236 .

A claimant is disabled for purposes of the SSA if she is unable to "engage in any substantial gainful activity by reason of any medically determinable physical or mental impairment [that] can be expected to result in death or which has lasted or can be expected to last for a continuous period of not less than 12 months." 42 U.S.C. § 423 (d)(1)(A). The Social Security Administration has outlined a "five-step, sequential evaluation process" to determine whether a claimant is disabled:

(1) whether the claimant is currently engaged in substantial gainful activity; (2) whether the claimant has a severe impairment or combination of impairments; (3) whether the impairment meets or equals the severity of the specified impairments in the Listing of Impairments; (4) based on a "residual functional capacity" assessment, whether the claimant can perform any of his or her past relevant work despite the impairment; and (5) whether there are significant numbers of jobs in the national economy that the claimant can perform given the claimant's residual functional capacity, age, education, and work experience.

McIntyre v. Colvin , 758 F.3d 146 , 150 (2d Cir. 2014) (citing Burgess v. Astrue , 537 F.3d 117 , 120 (2d Cir. 2008) ; 20 C.F.R. §§ 404.1520 (a)(4)(i)-(v), 416.920(a)(4)(i)-(v) ). The claimant bears the burden of proving her case at steps one through four. Id. The burden shifts to the Commissioner at step five. Id.

On October 22, 2012, the Commissioner denied Estrella's application. The Social Security Administration's Office of Disability Adjudication and Review subsequently granted Estrella's request for a hearing before an ALJ. Following a hearing in October 2013, and a second hearing in January 2014, the ALJ determined, in accordance with the five-step process, the following: (1) Estrella "ha[d] not engaged in substantial gainful activity since August 27, 2008," Admin. R. 25; (2) her mild lumbar and median nerve radiculopathy, mild to moderate bilateral carpal tunnel syndrome, diabetes, and depression were severe impairments because the conditions "impose[d] more than minimal limitations" on her ability to perform basic work activities, id. ; (3) she did not have an impairment or combination of impairments that met or equaled the severity of the specified impairments in the Listing of Impairments, id. ; (4) she had the residual functional capacity ("RFC") to perform light work, with certain limitations, and she could perform past relevant work as an administrative clerk, which would not be precluded by her RFC, id. at 27, 30 ; and (5) she could otherwise perform a "wide range of light level ... work," id. at 30 .

*95 The ALJ accordingly denied Estrella's application. Id.

The Office of Disability Adjudication and Review denied Estrella's request for review on May 14, 2015, at which point the ALJ's decision became the final decision of the Commissioner. On July 8, 2015, Estrella filed a complaint in the United States District Court for the Eastern District of New York, effectively appealing the Commissioner's decision. On September 6, 2017, the district court denied Estrella's motion for judgment on the pleadings and granted the Commissioner's cross-motion for the same. This appeal followed.

DISCUSSION

I. Standard of Review

On an appeal from the denial of disability benefits, "we focus on the administrative ruling rather than the district court's opinion." McIntyre , 758 F.3d at 149 (citation omitted). "We conduct a plenary review of the administrative record to determine if there is substantial evidence, considering the record as a whole, to support the Commissioner's decision and if the correct legal standards have been applied." Cichocki v. Astrue , 729 F.3d 172 , 175-76 (2d Cir. 2013) (per curiam) (quotation marks and citation omitted). Substantial evidence is evidence that "a reasonable mind might accept as adequate to support a conclusion." McIntyre , 758 F.3d at 149 (citation omitted). Although we do not require that "every conflict in a record be reconciled by the ALJ or the Secretary, ... we do [require] that the crucial factors in any determination ... be set forth with sufficient specificity to enable us to decide whether the determination is supported by substantial evidence." Ferraris v. Heckler , 728 F.2d 582 , 587 (2d Cir. 1984).

II. Treating Physician Rule

Estrella principally challenges the ALJ's handling of her treating psychiatrist's opinion. In October 2013, Dr. Felix Dron submitted a Medical Source Statement to the Office of Disability Adjudication and Review in which he opined that Estrella's "poor concentration, forgetful[ness], low stress tolerance, rage outbursts[,] ... depression[,] and anxiety" resulted in "marked" limitations in the areas of making work-related decisions and understanding, remembering, and carrying out detailed instructions. Admin. R. 507-08. Although Dr. Dron had treated Estrella for roughly five years, the ALJ assigned "little weight" to his opinion. Id. at 29. Estrella argues that, in doing so, the ALJ failed to comply with the procedural mandates of the so-called treating physician rule. We agree.

Social Security Administration regulations, as well as our precedent, mandate specific procedures that an ALJ must follow in determining the appropriate weight to assign a treating physician's opinion. First, the ALJ must decide whether the opinion is entitled to controlling weight. "[T]he opinion of a claimant's treating physician as to the nature and severity of [an] impairment is given 'controlling weight' so long as it 'is well-supported by medically acceptable clinical and laboratory diagnostic techniques and is not inconsistent with the other substantial evidence in [the] case record.' " Burgess , 537 F.3d at 128 (third brackets in original) (quoting 20 C.F.R. § 404.1527 (c)(2) ). Second, if the ALJ decides the opinion is not entitled to controlling weight, it must determine how much weight, if any, to give it. In doing so, it must "explicitly consider" the following, nonexclusive " Burgess factors": "(1) the frequen[cy], length, nature, and extent of treatment; (2) the amount of medical evidence supporting the opinion;

*96 (3) the consistency of the opinion with the remaining medical evidence; and (4) whether the physician is a specialist." Selian v. Astrue , 708 F.3d 409 , 418 (2d Cir. 2013) (per curiam) (citing Burgess , 537 F.3d at 129 (citing 20 C.F.R. § 404.1527 (c)(2) )). At both steps, the ALJ must "give good reasons in [its] notice of determination or decision for the weight [it gives the] treating source's [medical] opinion." Halloran v. Barnhart , 362 F.3d 28 , 32 (2d Cir. 2004) (per curiam) (quoting 20 C.F.R. § 404.1527 (c)(2) ).

An ALJ's failure to "explicitly" apply the Burgess factors when assigning weight at step two is a procedural error. Selian , 708 F.3d at 419-20 . If "the Commissioner has not [otherwise] provided 'good reasons' [for its weight assignment]," we are unable to conclude that the error was harmless and consequently remand for the ALJ to "comprehensively set forth [its] reasons." See Halloran , 362 F.3d at 33 . If, however, "a searching review of the record" assures us "that the substance of the treating physician rule was not traversed," we will affirm. See id. at 32 .

To begin, although substantial evidence supports the ALJ's decision at step one to assign less-than-controlling weight to Dr. Dron's opinion, the same is not true of its decision at step two to assign "little weight" thereto. In deciding, at step one, that Dr. Dron's opinion was not entitled to controlling weight, the ALJ pointed to two positive treatment notes from July and August 2012. The October 2013 Medical Source Statement, it explained,

[was] inconsistent with the evidence and [Dr. Dron's own] examination of [Estrella]. During [Dr. Dron's] examination of [Estrella], ... most of her mental status examinations were normal, she only had at most mild depression, she denied having hallucination [ sic ], and suicidal, homicidal ideations, and there was no evidence of cognitive limitations or psychosis. Most importantly, from September 12, 2012 through October 15, 2013, [Estrella's] [Global Assessment of Functioning] scores [had] been consistently at 70, which indicate[d] mild depression, not marked depression.

Admin. R. 29. In subsequently deciding, at step two, to assign "little weight" to Dr. Dron's opinion, the ALJ provided no additional reasoning. It thus failed to "explicitly consider" the first Burgess factor-"the frequen[cy], length, nature, and extent of [Dr. Dron's] treatment"-before weighing the value of the opinion. See Selian , 708 F.3d at 418 . Nowhere in the ALJ's decision, for example, is the fact that Dr. Dron treated Estrella from 2004 to 2006, and again from 2010 to 2013. Nor did the ALJ consider that Dr. Dron prescribed various psychotropic medications over the course of his treatment, in addition to providing monthly psychotherapy sessions.

Because the ALJ procedurally erred, the question becomes whether "a searching review of the record ... assure[s us] ... that the substance of the ... rule was not traversed"- i.e. , whether the record otherwise provides "good reasons" for assigning "little weight" to Dr. Dron's opinion. See Halloran , 362 F.3d at 32 . It does not.

First, a number of treatment notes not cited by the ALJ reflect a more serious impairment than the ALJ acknowledged. For example, in January 2012, prior to the only two referenced notes, Dr. Dron observed that Estrella

[was] inattentive[,] ... ha[d] a short attention span[,] ... often [did] not seem to be listening when spoken to directly[,] ... often [did] not follow through on instructions[,] and fail[ed] to finish tasks. She [was] disorganized ... [and avoided t]asks that require sustained mental effort *97 .... She often los[t] things necessary for tasks or activities[,] ... [and was] easily distracted ... [and] forgetful.

Admin. R. 456. In June 2013, subsequent to the two referenced notes and closer in time to the ALJ's decision, Dr. Dron observed that Estrella's

anxiety symptoms continue[d] .... [She] continue[d] to be inattentive[,] ... [and did] not seem to be listening when spoken to directly .... Tasks that require[d] sustained mental effort [were] still being avoided .... [Estrella] continue[d] to often lose things necessary for tasks or activities[,] ... [was] still as forgetful as previously[,] ... [and showed] signs of anxiety.

Id. at 434.

The ALJ made no attempt to "reconcile" or "grapple with" the apparent longitudinal inconsistencies in Estrella's mental health-one of the motivations behind Burgess 's procedural requirement of explicit consideration of "the frequen[cy], length, nature, and extent of [a physician's] treatment." See Selian , 708 F.3d at 418-19 . This failure is especially relevant here because the first Burgess factor, and therefore evidence supporting its satisfaction, is of heightened importance in the context of Estrella's claimed impairment: depression. "Cycles of improvement and debilitating symptoms [of mental illness] are a common occurrence, and in such circumstances it is error for an ALJ to pick out a few isolated instances of improvement over a period of months or years and to treat them as a basis for concluding a claimant is capable of working." Garrison v. Colvin , 759 F.3d 995 , 1017 (9th Cir. 2014) ; see also Bauer v. Astrue , 532 F.3d 606 , 609 (7th Cir. 2008) ("A person who has a chronic disease, whether physical or psychiatric, and is under continuous treatment for it with heavy drugs, is likely to have better days and worse days .... Suppose that half the time she is well enough that she could work, and half the time she is not. Then she could not hold down a full-time job."). When viewed alongside the evidence of the apparently cyclical nature of Estrella's depression, the ALJ's two cherry-picked treatment notes do not provide "good reasons" for minimalizing Dr. Dron's opinion.

Second, Estrella's Global Assessment of Functioning ("GAF") scores, which the ALJ concluded "indicated mild depression, not marked depression," Admin. R. 29, do not provide good reasons for assigning little weight to Dr. Dron's opinion. 1 The Social Security Administration has explained that "[u]nless [a] clinician clearly explains the reasons behind his or her GAF rating, and the period to which the rating applies, it does not provide a reliable longitudinal picture of the claimant's mental functioning for a disability analysis." U.S. Soc. Sec. Admin., Office of Disability Programs, AM-13066, Global Assessment of Functioning (GAF) Evidence in Disability Adjudication (Oct. 14, 2014). Furthermore, "[u]nless the GAF rating is well supported and consistent with other evidence in the file, it is entitled to little weight under our rules." Id. Because Estrella's GAF scores were bereft of any explanation of Dr. Dron's reasoning, and because they are unsupported by Dr. Dron's other conclusions as to the severity *98 of Estrella's depression, they do not provide "good reasons" for assigning little weight to Dr. Dron's opinion.

Finally, the opinion of Dr. Christopher Flach, a one-time consultative psychologist, similarly does not provide a good reason for diminishing Dr. Dron's opinion. After examining Estrella once in September 2012, Dr. Flach opined that Estrella had mild problems maintaining attention and concentration but could follow and understand simple directions and instructions, maintain a regular schedule, learn new tasks, perform complex tasks independently, and make appropriate decisions. The ALJ assigned "substantial weight" to Dr. Flach's opinion because it "[was] consistent with [Dr. Dron's] mental treatment notes, which [did] not show marked limitations." Admin. R. 29.

We have frequently "cautioned that ALJs should not rely heavily on the findings of consultative physicians after a single examination." Selian , 708 F.3d at 419 . This concern is even more pronounced in the context of mental illness where, as discussed above, a one-time snapshot of a claimant's status may not be indicative of her longitudinal mental health. Because the ALJ made the same error with respect to Dr. Flach as it did with respect to Dr. Dron-namely, relying on his opinion without "reconcil[ing]" or "grappl[ing] with" Estrella's fluctuating state of mental health-we are not assured that, had the ALJ complied with the procedural mandates of the treating physician rule, it would still have assigned significant weight to Dr. Flach's opinion. The opinion of Dr. Flach, in other words, similarly does not provide a "good reason" for minimizing that of Dr. Dron.

In light of the ALJ's failure to "explicitly consider" the first Burgess factor before assigning "little weight" to the opinion of Estrella's treating psychiatrist, and the lack of other "good reasons" to support that decision, we conclude that the ALJ traversed the substance of the treating physician rule. 2 Accordingly, we remand to the ALJ for reconsideration of Estrella's claim for disability benefits consistent with the procedural mandates of the SSA and this Court. 3 On remand, the ALJ should apply all four Burgess factors in determining the appropriate weight to accord to Dr. Dron's opinion.

CONCLUSION

For the foregoing reasons, the judgment of the district court is VACATED, and the case is REMANDED to the Commissioner for further proceedings consistent with this opinion.